Read Me for the code for �Multivariate Forecast Evaluation and Rationality Testing� by Ivana Komunjer and Michael Owyang

The folder includes 13 files in the main folder. The main file that produces the empirical result is multiloss.  The file that produces the Monte Carlo results is multiloss_mc which calls multiloss.
Running the code requires the statistics and optimization toolboxes for Matlab
The real-time data used to make the instruments and evaluate the forecasts can be found in the folder �final real-time data�. The forecast data is proprietary and is not included.
The folder �figure code� makes figures 2 and 3 in the paper. Figure 1 was created directly in Scientific Workplace.
